where can i get the pose result?

[smart4654154]

you let the network predict the camera poses for all input images while also learning the varying scale factors across different scenes.

so where can I get the predicted pose result?
thank you.
What I mean is, where can the predicted pose be located and output it?
I want to generate results similar to colmap matching, such as bin or blender JSON files
So I can use other algorithms for comparative testing. Because other algorithms require poses